Hints & Tips

Q: Can I use the Virtual Wafer Fab to tune my simulations? For example, can I perform "split-lot" simulation using a diffusion coefficient as a variable?

A: The Virtual Wafer Fab provides a very convenient way to emulate split lots using simulation software. A typical VWF experiment is to examine the effect of a particular implant dose on the device breakdown characteristics. However, the specification of which quantities to vary in a VWF "split lot" experiment is left entirely to the user, who can use any combination of process variables and model parameters. It is therefore possible to tune simulations by defining experiments that vary diffusion coefficients, segregation coefficients, or other model parameters.

Defining the split points in any input deck is a simple mouse-driven operation. Figure 1 shows the windows used to select the DIP.0 boron diffusion coefficient. The DeckBuild EXTRACT statement allows a wide range of parameters to be used as targets for tuning experiments. In particular electrical variables such as threshold voltage and sheet resistance can be extracted from ATHENA/SSUPREM4 simulations. Figure 2 shows the response of the sheet resistance of a p-type layer to a boron diffusion coefficient in ATHENA/SSUPREM4.

 

 


Figure 1.




Figure 2.

 

Q: Can I experiment with grid density in VWF?

A: Users concerned with ensuring sufficient grid density for a particular simulation can define VWF experiments that vary the grid density. Figure 3 shows results from a VWF experiment that characterizes the breakdown voltage of a diode as a function of grid density. A combined ATHENA/ATLAS input deck for the diode fabrication and breakdown determination was made. Splits were done on the number of grid layers in the expitaxial layer. For 130 or more grid points the breakdown voltage varies by less than 2%. At low grid densities the breakdown voltage varies by up to 20% from the limiting value.

Figure 3.